Re: [Qemu-block] [Qemu-devel] [PATCH] mirror: hold aio_context before bdrv_drain

[Bin Wu]

On 2015/4/1 16:19, Fam Zheng wrote:
> On Wed, 04/01 12:42, Bin Wu wrote:
>> From: Bin Wu <address@hidden>
> 
> What's the issue are you fixing? I think the coroutine already is running in
> the AioContext of bs.
> 
> Fam
> 
In the current implementation of bdrv_drain, it should be placed in a critical
section as suggested in the comments above the function: "Note that unlike
bdrv_drain_all(), the caller must hold the BlockDriverState AioContext".

However, the mirror coroutine starting with mirror_run doesn't do this. I just
found qmp_drive_mirror protects the AioCentext, but it is out of the scope of
the mirror coroutine.

Bin

>>
>> Signed-off-by: Bin Wu <address@hidden>
>> ---
>>  block/mirror.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/block/mirror.c b/block/mirror.c
>> index 4056164..08372df 100644
>> --- a/block/mirror.c
>> +++ b/block/mirror.c
>> @@ -530,7 +530,9 @@ static void coroutine_fn mirror_run(void *opaque)
>>               * mirror_populate runs.
>>               */
>>              trace_mirror_before_drain(s, cnt);
>> +            aio_context_acquire(bdrv_get_aio_context(bs));
>>              bdrv_drain(bs);
>> +            aio_context_release(bdrv_get_aio_context(bs));
>>              cnt = bdrv_get_dirty_count(bs, s->dirty_bitmap);
>>          }
>>  
>> -- 
>> 1.7.12.4
>>
>>
>>
> 
> .
>